Fraud Detection System
Detection pipeline over 284,807 transactions using a stacked ensemble plus an ARIMA anomaly-detection layer for time-series signal.
I turn raw data into decisions - and increasingly, into AI systems that reason.
I'm a Computer Science graduate from Bennett University (CGPA 8.14), working across data analytics, machine learning and generative AI. My background spans a Decision Scientist Trainee stint at Mu Sigma and a Data Analyst internship at NTPC, plus a Springer-published research paper on deep transfer learning for EEG classification.
I like problems that sit between analysis and engineering: fraud-detection pipelines, RAG chatbots, credit-risk models, and Power BI dashboards that people actually use.
Detection pipeline over 284,807 transactions using a stacked ensemble plus an ARIMA anomaly-detection layer for time-series signal.
RAG assistant built on LangGraph with a FAISS vector store, multi-LLM support (Ollama / Gemini / OpenAI), SQLite persistence and auto-generated chat titles.
Interactive Power BI dashboard analyzing customer behavior, credit usage and income trends, backed by SQL and Python EDA.
Exploratory data analysis on BlinkIt sales data in Python to understand purchase patterns and improve customer experience.
Springer-published research (SmartCom 2026, Vol. 1994) on deep transfer learning for EEG-based classification. DOI: 10.1007/978-3-032-27160-0_42.
Open to Data Analyst, Data Scientist, Decision Scientist and GenAI Engineer roles across Delhi NCR, Noida, Bangalore and remote. Let's talk.